Method, system and related device for retrieving image data

Abstract

This invention discloses a method, system, and related equipment for retrieving image data. The method first constructs a medical image database containing multiple target examination data sets, each including at least one image sequence. Second, it extracts textual description information and image content information from each image sequence and establishes an image sequence index based on these information. Further, based on the image sequence index corresponding to at least one image sequence included in the same target examination data set, it aggregates information according to preset aggregation rules to generate a target examination data index. Finally, in response to a received retrieval request, it performs a retrieval in the medical image database based on the image sequence index and / or the target examination data index, and outputs retrieval results that satisfy the retrieval request.

Description

Technical Field

[0001] This application belongs to the field of data retrieval and medical information technology, specifically relating to an image data retrieval method, an image data retrieval system, an electronic device, and a computer-readable storage medium. Background Technology

[0002] With the advancement of medical informatization, medical data is experiencing explosive growth. Multimodal medical imaging data, such as CT and MRI, along with diagnostic reports, constitute core medical resources. Accurate retrieval of image datasets that meet research needs has become a crucial foundation for medical research. Currently, the mainstream method for retrieving and classifying image data involves extracting and matching DICOM (Digital Imaging and Communications in Medicine) field information (such as patient information and examination parameters) and key text content from diagnostic reports. This method is widely used in clinical data management and medical research.

[0003] However, this mainstream retrieval technology has significant shortcomings, which severely restrict the efficiency of data utilization and the quality of scientific research: On the one hand, due to factors such as equipment failure and operational errors, medical digital imaging and communication standard fields are prone to missing or incorrect information, and image diagnosis reports are mostly comprehensive descriptions of multiple sequences of images from a single examination, which cannot accurately match the retrieval needs of single-sequence images, directly leading to insufficient accuracy in retrieval and classification; on the other hand, this type of technology can only mine textual information and cannot capture image content features such as lesion morphology and density, resulting in a single dimension of information utilization, which is difficult to meet researchers' in-depth needs for studying the characteristics of disease images and provides limited support.

[0004] In summary, current medical image data retrieval technologies suffer from several core problems, including insufficient reliability of standard field information in medical digital imaging and communication, lack of accuracy in diagnostic report annotations, and limited utilization of retrieval information dimensions. These problems directly affect the efficiency and quality of research data screening. Therefore, developing medical image data retrieval technologies that combine accuracy with multi-dimensional information mining capabilities has become a critical issue that urgently needs to be addressed in the field of medical informatics. Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method, system, and related equipment for retrieving image data.

[0006] In a first aspect, embodiments of this application provide a method for retrieving image data, including: Construct a medical imaging database, which contains multiple target examination data, each of which includes at least one image sequence; Extract text description information and image content information for each image sequence, and build an image sequence index based on the text description information and image content information; Based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to preset aggregation rules to generate a target inspection data index. In response to receiving a search request, the system performs a search in the medical imaging database based on the image sequence index and / or the target examination data index, and outputs search results that satisfy the search request.

[0007] In some embodiments, text description information and image content information of each image sequence are extracted, and an image sequence index is built based on the text description information and image content information, including: Obtain image reports associated with image sequences; A text information index is constructed based on metadata field information from image reports and / or image sequences; An image content index is constructed based on the image content information contained in the image sequence; An image sequence index is generated based on text information indexing and image content indexing.

[0008] In some embodiments, a text information index is constructed based on metadata field information of image reports and / or image sequences, including: Extract the first text information from the metadata fields of the image sequence; Extracting second text information from image reports associated with image sequences; A text information index is constructed based on the first text information and / or the second text information.

[0009] In some embodiments, an image content index is constructed based on the image content information contained in the image sequence, including: Image content information is obtained by recognizing image sequences using an image content recognition model. Generate corresponding index items based on the type of image content information; The index items are combined to form an image content index for the image sequence.

[0010] In some embodiments, the image content information includes classification information and severity information; Generate corresponding index entries based on the type of image content information, including: Based on the classification information, a classification index item is generated. The classification information is used to characterize whether there is image content of a preset category in the image sequence. Severity index entries are generated based on severity information, which is used to characterize the degree of presence of specific image content in an image sequence.

[0011] In some embodiments, the aggregation rule includes at least one of the following: For the classification index items, classify index items of the same type in the same target inspection data and merge them. For severity-based index items, aggregation calculations are performed based on the quantified scores corresponding to the severity-based index items; The text information index is integrated and processed.

[0012] In some embodiments, based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to a preset aggregation rule to generate a target inspection data index, including: Merge the judgment classification index items of the same type in the same target inspection data to generate the first aggregate index; Aggregate the quantified scores corresponding to severity-type index items of the same type in the same target inspection data to generate a second aggregated index; By integrating the text information indexes corresponding to the same target inspection data, a third aggregated index is generated; A target inspection data index is generated based on the first, second, and third aggregated indexes.

[0013] Secondly, embodiments of this application provide an image data retrieval system, comprising: The database construction module is configured to build a medical image database containing multiple target examination data, each of which includes at least one image sequence. The sequence index building module is configured to extract text description information and image content information for each image sequence, and build an image sequence index based on the text description information and image content information; The data index building module is configured to aggregate information according to preset aggregation rules based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, and generate the target inspection data index. The data retrieval module is configured to, in response to receiving a retrieval request, perform a retrieval in the medical imaging database based on the image sequence index and / or the target examination data index, and output retrieval results that satisfy the retrieval request.

[0014] Thirdly, embodiments of this application provide an electronic device, including: a processor and a memory, wherein the memory stores a program or instructions that can run on the processor, and when the program or instructions are executed by the processor, they implement the image data retrieval method as described in the first aspect.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium on which a program or instructions are stored, and when the program or instructions are executed by a processor, the image data retrieval method as described in the first aspect is implemented.

[0016] The image data retrieval method, system, and related equipment provided in this application first construct a medical image database containing multiple target examination data, each including at least one image sequence. Second, textual description information and image content information are extracted from each image sequence, and an image sequence index is established based on these information. Further, based on the image sequence index corresponding to at least one image sequence included in the same target examination data, information is aggregated according to preset aggregation rules to generate a target examination data index. Finally, in response to a received retrieval request, a retrieval is performed in the medical image database based on the image sequence index and / or the target examination data index, and retrieval results satisfying the retrieval request are output. This application categorizes image content information from multiple dimensions and conducts retrieval based on the textual information and specific image content information of the medical image data. This retrieval method not only eliminates the influence of missing or incorrect image data field information but also supports joint retrieval of image and text information, and provides multi-type feature indexes and retrieval functions related to image content. By accurately retrieving patient examination data, image sequences, and their text reports, the granularity of the retrieval is improved, thereby achieving accurate retrieval and aggregation of medical image data, and more efficiently completing the retrieval of target image data and the construction of datasets.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart illustrating an image data retrieval method according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of an image data retrieval system according to an embodiment of this application; Figure 3 This is a block diagram of an electronic device according to some embodiments of this application.

[0019] Explanation of reference numerals in the attached figures: image data retrieval system 200, database construction module 201, sequence index construction module 202, data index construction module 203, data retrieval module 204, processor 310, memory 320, input / output interface 330, communication interface 340, and bus 350. Detailed Implementation

[0020] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0021] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0022] As described in the background section, current medical image data retrieval technologies suffer from problems such as insufficient reliability of standard field information in medical digital imaging and communication, lack of accuracy in diagnostic report annotations, and limited utilization of retrieval information dimensions. This application provides a method, system, and related equipment for retrieving image data. First, a medical image database containing multiple target examination data is constructed. Then, metadata and image report text information of each image sequence are extracted to build a text information index. Simultaneously, image content information is obtained through an image content recognition model to generate an image content index, thereby establishing an image sequence index. The system merges and integrates various indexes under the same target examination data according to preset aggregation rules to generate a corresponding target examination data index. Upon receiving a retrieval request, the system performs a search in the database based on the above two types of indexes and outputs the required retrieval results.

[0023] refer to Figure 1 This is a flowchart illustrating an image data retrieval method in an embodiment of this application.

[0024] like Figure 1 As shown, the image data retrieval methods include: Step S101: Construct a medical image database containing multiple target examination data, each target examination data including at least one image sequence.

[0025] In practice, a medical imaging database is constructed by collecting medical imaging data from multiple modalities and their corresponding imaging diagnostic reports. This database stores patient examination data, including imaging report data. Each examination record represents a single patient's examination and contains at least one image sequence.

[0026] Step S102: Extract the text description information and image content information of each image sequence, and build an image sequence index based on the text description information and image content information.

[0027] In practical implementation, patient examination data in the medical imaging database can be used to build an image sequence index based on text description information and image content information in the image sequence. Text description information can be extracted directly from the metadata fields of the image sequence (first text information) or by first obtaining the image report associated with the image sequence and then extracting the second text information from the image report. Ultimately, based on one or both of the first and second text information, a complete text description of the image sequence can be constructed for subsequent text information indexing. Image content information is obtained by recognizing the target image sequence using an image content recognition model. The recognized image content information is divided into two categories: one is classification information, used to characterize whether image content of a preset category exists in the image sequence; the other is severity information, used to characterize the degree of presence of specific image content in the image sequence. These two types of information together constitute complete image content information, which can be used to generate corresponding index entries to construct the image content index.

[0028] Step S103: Based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to a preset aggregation rule to generate a target inspection data index.

[0029] In the specific implementation process, based on the established image sequence index, the system performs information aggregation processing according to preset aggregation rules, ultimately generating a higher-level examination data index. The system merges the categorized index items into a union by type to obtain the first aggregated index; it performs comprehensive calculations on the severity index items based on quantitative scores to obtain the second aggregated index; and it performs deduplication and integration processing on the text information index to obtain the third aggregated index. The first, second, and third aggregated indexes are then combined to form the complete examination data index for this image examination, which provides support for efficient retrieval in the database later.

[0030] Step S104: In response to receiving a search request, perform a search in the medical imaging database based on the image sequence index and / or the target examination data index, and output the search results that satisfy the search request.

[0031] In practice, users can directly input the image content or text information they wish to find at the search portal to obtain patient imaging examination data or image sequence data that meets their requirements. After the user inputs search information, such as the actual scanned area, the severity of artifacts, or the severity of noise, the system will rely on the examination data index to perform a data retrieval in the medical imaging database, aggregate the search results into similar data sets, and finally output this set. This achieves precise matching of text and image content information, enabling retrieval of similar data for specific areas or with certain characteristics.

[0032] In some embodiments, text description information and image content information of each image sequence are extracted, and an image sequence index is built based on the text description information and image content information, including: obtaining an image report associated with the image sequence; constructing a text information index based on the image report and / or metadata field information of the image sequence; constructing an image content index based on the image content information contained in the image sequence; and generating an image sequence index based on the text information index and the image content index.

[0033] In its implementation, the system first extracts textual descriptions and metadata fields conforming to medical digital imaging and communication standards from the image sequence. Then, it combines this with the textual content of the image report associated with the image sequence and the extracted metadata fields to construct a textual information index. Simultaneously, the system can use a pre-trained machine learning model to perform analysis and recognition operations on the image sequence, extracting image content information. Based on the type of extracted image content information, corresponding index items are generated. For categorization information indicating the presence of image content of a preset category, the system generates categorization index items. For severity information describing the degree of presence of a specific image type, the system generates severity index items. The system combines these generated index items to ultimately form a complete image content index for the image sequence.

[0034] After constructing the independent text information index and image content index, the system merges and correlates these two indexes to form a unified and comprehensive image sequence index. This index fully covers the text descriptions from image reports and image sequences, as well as the image content features identified through machine learning models, enabling the image sequence to be characterized from both textual and visual content dimensions. This comprehensive image sequence index then lays the data foundation for subsequent aggregation by inspection level to generate higher-level inspection data indexes.

[0035] In some embodiments, constructing a text information index based on metadata field information of image reports and / or image sequences includes: extracting first text information from metadata field information of image sequences; extracting second text information from image reports associated with image sequences; and constructing a text information index based on the first text information and / or the second text information.

[0036] In its implementation, the system collects field information from each image sequence in the patient's examination data and extracts the text content from the corresponding imaging reports to construct a text information index. Specifically, for each patient's examination data, the system extracts metadata field information conforming to medical digital imaging and communication standards from each image sequence, and simultaneously collects the text content from the patient's imaging report. This results in a text information index containing information such as age, gender, modality, number of sequences, number of images, image description, and diagnostic conclusion. The metadata field information conforming to medical digital imaging and communication standards extracted from each image sequence constitutes the first text information, while the text content extracted from the imaging report associated with that image sequence constitutes the second text information. Based on this text information index, users can input relevant text information to quickly retrieve matching medical imaging data and text reports.

[0037] In some embodiments, constructing an image content index based on image content information contained in an image sequence includes: identifying the image sequence using an image content recognition model to obtain image content information; generating corresponding index entries based on the type of image content information; and combining the index entries to form an image content index of the image sequence.

[0038] In practical implementation, the image content recognition function of image sequences can be achieved by a single deep learning model with comprehensive functions or multiple independent deep learning models. The system first builds a deep neural network model, which can use a ResNet50 classification model as its main structure. Then, preprocessed medical image data, after normalization and other operations, is input into this classification model. Simultaneously, multiple classification labels are set as the model's learning targets to complete the training of the classification network. After training, the image content recognition model has the ability to identify the scanned areas in the image. In the model inference stage, the system inputs the preprocessed image into the trained image content recognition model, obtaining multiple classification prediction results distributed in the range of 0 to 1. These classification prediction results serve as index items. The system selects a preset threshold, for example, 0.5. Classification results with a prediction probability greater than or equal to 0.5 are determined to indicate the existence of a corresponding target, while classification results with a prediction probability less than 0.5 are determined to indicate the absence of a corresponding target.

[0039] One feasible approach is to establish a multi-class prediction model, which can be used to identify whether scanned areas such as the head, neck, and lungs exist in image sequences. The specific operation process is as follows: The system first constructs an image part recognition dataset, and completes multiple information annotations for each data point, including whether it contains a head, chest, etc. The annotation rule is to use 1 to indicate the presence of the corresponding part and 0 to indicate its absence. For index items such as scanned areas, which cover multiple types of image content, after identifying and indexing each image sequence in each patient's examination data, the system also needs to summarize the scanned area information corresponding to each sequence. The summarized information is then used as the scanned area index for that patient's examination data. For example, if a patient's examination data contains multiple image sequences, one of which scans the head and another of which scans the abdomen, then the scanned area index for that patient's examination data will cover both head and abdomen information. Users can retrieve the patient's examination data by entering any relevant search keyword.

[0040] In some embodiments, the image content information includes classification information and severity information; generating corresponding index entries based on the type of image content information includes: generating classification index entries based on classification information, wherein the classification information is used to characterize whether there is image content of a preset category in the image sequence; and generating severity index entries based on severity information, wherein the severity information is used to characterize the degree of existence of specific image content in the image sequence.

[0041] In practical implementation, the classification information is used to characterize whether a preset category of image content exists in the image sequence. This can include the examination site, whether angiography is present, and whether an implant is present. The system generates classification index items based on the classification information. The labeling rule for this type of information is 0 to indicate the absence of corresponding content and 1 to indicate the presence of corresponding content. For example, if an image sequence contains angiography-related content, the classification information for whether angiography is present in that sequence is labeled as 1; if the sequence does not contain angiography-related content, the classification information for whether angiography is present is labeled as 0. If any one of the multiple image sequences in a patient's examination data contains a certain type of image content, then the patient's examination data is directly marked as containing that type of image content. For example, if a patient's examination data contains multiple image sequences, one of which contains an artificial joint, while the other image sequences do not, then this patient's examination data is marked as containing an implant.

[0042] Severity information is used to characterize the presence degree of a specific image type in an image sequence, and can include the severity of artifacts, noise severity, etc. The system generates severity index entries based on this information, and uses a differentiated labeling rule for this type of information: 0 represents a slight severity, and 1 represents a severe severity, with values ​​in the range of 0 to 1 corresponding to different degrees of severity. For example, if an image sequence contains only a few artifacts, the labeling value for the severity information of artifact severity can be set to 0.2; if the noise effect of the sequence is significant, the labeling value for the severity information of noise severity can be set to 0.7.

[0043] Furthermore, if an implant and significant noise are present in the image sequence corresponding to a patient's examination data, when a user retrieves the patient's examination data, they only need to select any one of the following as search criteria: patient name, presence of implant, or presence of significant noise, to retrieve the corresponding examination data.

[0044] In some embodiments, the aggregation rules include at least one of the following: for judgment classification index items, merging judgment classification index items of the same type in the same target inspection data; for severity index items, performing aggregation calculation based on the quantitative score corresponding to the severity index item; and for text information index, performing integration processing.

[0045] In practical implementation, for multiple classification index items from different image sequences under the same examination, the system merges index items of the same type, ultimately generating a unique set of index items that covers all identified categories under that examination. For multi-type image content, the scanned body part identification results for each image sequence are divided into two categories: single-body and multi-body. When the scanned body part is single, the system only establishes an information index for that body part; when the scanned body part is multi-body, the system establishes information indexes for all four body parts. For example, if a CT scan sequence involves multiple body parts (head, neck, chest, abdomen), the system establishes indexes for four body parts (head, neck, chest, abdomen) based on that image data. The retrieval end can use AND, OR, NOT, and other search conditions and methods to achieve precise or fuzzy retrieval of any number of keywords among the four body parts, thereby obtaining image sequences containing specific scanned body parts. For judgmental image content such as whether imaging is present or whether implants are present, the system can directly establish an index to determine whether the corresponding image content exists. Deduplication and summarization can be performed at the inspection level to avoid information redundancy caused by multiple sequences identifying the same category within the same inspection. Finally, a concise and complete list of categories is generated to represent the overall image content characteristics of the inspection.

[0046] For severity-based image content indexing, the system establishes an index based on the numerical value or range of severity corresponding to the image sequence. Aggregation of severity-based index items is performed based on the quantified score corresponding to each index item. This quantified score objectively characterizes the quantitative assessment result of the image sequence for a specific severity type. For multiple index items belonging to the same severity type, the system uses preset mathematical rules to comprehensively calculate their quantified scores, thereby generating a single aggregated value that represents the overall severity of the entire examination on that feature. For image content describing severity, such as artifact severity and noise severity, the system uses both the highest and lowest severity values ​​in the image sequence as index information for the patient's examination image data. The search terminal can choose to search for the patient's examination image data by the highest or lowest severity of the image sequence content. For example, a patient's examination image data may contain multiple image sequences, one of which contains the lowest severity respiratory artifact value. Another image sequence contains the highest severity values ​​for respiratory artifacts. The system then records both artifact severity values ​​in the patient's examination data index. This process combines quantitative assessments of multiple sequences into a single, representative comprehensive assessment, resulting in a concise, quantifiable severity index at the examination level for efficient retrieval and comparison.

[0047] The system integrates all text information indexes associated with a single examination. These indexes originate from sources such as image report texts, medical digital imaging, and communication standard metadata fields. The core method of integration is to merge, deduplicate, and fuse text index information from different sources to generate a unified, complete, and non-redundant examination-level text index. Specifically, the system first aggregates the text information indexes corresponding to image sequences belonging to the same examination, then eliminates completely duplicated or semantically repetitive index entries, and finally merges and associates relevant information to form a structured and easily searchable text index set. Through this series of processes, the system extracts text information scattered across multiple sequences and reports into a single, high-quality text summary index. This operation avoids information fragmentation and ensures efficiency and accuracy in text retrieval at the examination level.

[0048] In some embodiments, based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to a preset aggregation rule to generate a target inspection data index, including: merging judgment classification index items of the same type in the same target inspection data to generate a first aggregated index; aggregating the quantitative scores corresponding to the severity index items of the same type in the same target inspection data to generate a second aggregated index; integrating the text information indexes corresponding to the same target inspection data to generate a third aggregated index; and generating a target inspection data index based on the first aggregated index, the second aggregated index, and the third aggregated index.

[0049] In practice, the system merges the indexes of the same type from all image sequences under the same examination to generate a first aggregated index, which represents the preset image content categories present in that examination. For categorized searches such as scan site, presence of contrast agents, and presence of implants, the search terminal only needs to input the corresponding search conditions to perform the search operation. For example, inputting "male + head + presence of implants + image sequence" will retrieve all image sequences that meet the criteria of "male + head + presence of implants" from the database. Inputting "male + head + presence of implants + patient examination data" will retrieve all patient examination data that meet the search criteria of "male + head + presence of implants" from the database.

[0050] Similarly, the system aggregates severity-type index items of the same type across all image sequences under the same examination according to their corresponding quantified scores and preset mathematical rules. These preset mathematical rules can select either the maximum value or the average value. The resulting aggregated index reflects the overall severity of the examination on specific image features. For information retrieval reflecting severity, the retrieval terminal needs to input a severity value range along with the specified search criteria. For example, when searching for image sequences that meet a certain artifact severity level, the retrieval terminal needs to select the artifact severity search criteria and input the minimum search value. and maximum search value The numerical interval satisfies 0≤ ≤ ≤1, after inputting the value, all data within the specific artifact severity range will be retrieved. When retrieving patient examination data that meets a certain artifact severity level, the search terminal needs to select the artifact severity search criteria and set the maximum artifact severity value to select. Or minimum artifact level value As a basis for the search, a minimum search value is then set. and maximum search value As the range for artifact severity retrieval, this numerical interval satisfies 0 ≤ ≤ ≤1, once set, all data within a specific artifact severity range can be obtained.

[0051] The system merges, deduplicates, and fuses all relevant text information indexes under the same inspection. These text information indexes originate from image reports and metadata fields of medical digital imaging and communication standards. After processing, a third aggregated index is generated.

[0052] Finally, after generating the first, second, and third aggregated indexes respectively, the system merges and correlates these three to form the final inspection data index for this inspection. This inspection data index is structured data integrating image classification, severity quantification, and text information, providing a unified data foundation for subsequent efficient, multi-dimensional, and accurate retrieval.

[0053] In practical applications, the technical solution described in this application can be integrated into a medical imaging cloud platform as a data quality control module. By deploying the index construction and retrieval methods described in this solution, the platform can automatically parse text information and recognize image content for each patient examination data uploaded to the platform, generating structured index information. When the platform needs to perform data quality control, the system can perform retrieval and filtering based on preset quality control rules (such as filtering image sequences containing severe artifacts, marking examination data where the scanned area is inconsistent with the report, etc.), quickly locating data that does not meet quality requirements, assisting administrators in data cleaning and annotation, thereby effectively improving the overall quality and application efficiency of the platform's data. Furthermore, this function can also provide clinical research users with precise data filtering services based on image content features, supporting the rapid construction of high-quality research datasets.

[0054] It is understood that before using the technical solutions disclosed in the various embodiments of this application, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this application in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0055] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this application's technical solution, based on the prompt message.

[0056] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0057] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this application. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this application.

[0058] In summary, the image data retrieval method proposed in this application first constructs a medical image database containing multiple target examination data, each including at least one image sequence. Second, it extracts textual description information and image content information for each image sequence and establishes an image sequence index based on these information. Further, based on the image sequence index corresponding to at least one image sequence included in the same target examination data, it aggregates information according to preset aggregation rules to generate a target examination data index. Finally, in response to a received retrieval request, it performs a retrieval in the medical image database based on the image sequence index and / or the target examination data index, and outputs retrieval results that satisfy the retrieval request. This application divides image content information from multiple dimensions and conducts retrieval based on the textual information and specific image content information of the medical image data. This retrieval method not only eliminates the influence of missing or incorrect image data field information but also supports joint retrieval of image and text information and provides multi-type feature indexes and retrieval functions related to image content. Based on this, the system can accurately retrieve patient examination data, image sequences, and their text reports, improving the fineness of the retrieval granularity, thereby achieving accurate retrieval and collection of medical imaging data, and more efficiently completing the retrieval of target image data and the construction of datasets.

[0059] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described above.

[0060] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0061] Corresponding to the above embodiments, this application also proposes an image data retrieval system.

[0062] refer to Figure 2 The present application provides a schematic diagram of the structure of an image data retrieval system.

[0063] This application provides an image data retrieval system 200, including: Database construction module 201 is configured to construct a medical image database containing multiple target examination data, each target examination data including at least one image sequence. The sequence index building module 202 is configured to extract text description information and image content information for each image sequence, and build an image sequence index based on the text description information and image content information; The data index building module 203 is configured to aggregate information according to preset aggregation rules based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, and generate a target inspection data index. The data retrieval module 204 is configured to, in response to receiving a retrieval request, perform a retrieval in the medical imaging database based on the image sequence index and / or the target examination data index, and output retrieval results that satisfy the retrieval request.

[0064] Optionally, the sequence index building module 202 is also configured as follows: Extract text description information and image content information for each image sequence, and build an image sequence index based on the text description information and image content information, including: Obtain image reports associated with image sequences; A text information index is constructed based on metadata field information from image reports and / or image sequences; An image content index is constructed based on the image content information contained in the image sequence; An image sequence index is generated based on text information indexing and image content indexing.

[0065] Optionally, the sequence index building module 202 is also configured to: build a text information index based on metadata field information of image reports and / or image sequences, including: Extract the first text information from the metadata fields of the image sequence; Extracting second text information from image reports associated with image sequences; A text information index is constructed based on the first text information and / or the second text information.

[0066] Optionally, the sequence index building module 202 is also configured as follows: An image content index is constructed based on the image content information contained in the image sequence, including: Image content information is obtained by recognizing image sequences using an image content recognition model. Generate corresponding index items based on the type of image content information; The index items are combined to form an image content index for the image sequence.

[0067] Optionally, the sequence index building module 202 is also configured as follows: Image content information includes classification information and severity information; Generate corresponding index entries based on the type of image content information, including: Based on the classification information, a classification index item is generated. The classification information is used to characterize whether there is image content of a preset category in the image sequence. Severity index entries are generated based on severity information, which is used to characterize the degree of presence of specific image content in an image sequence.

[0068] Optionally, the data index building module 203 is also configured as follows: Aggregation rules include at least one of the following: For the classification index items, classify index items of the same type in the same target inspection data and merge them. For severity-based index items, aggregation calculations are performed based on the quantified scores corresponding to the severity-based index items; The text information index is integrated and processed.

[0069] Optionally, the data index building module 203 is also configured as follows: Based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to preset aggregation rules to generate a target inspection data index, including: Merge the judgment classification index items of the same type in the same target inspection data to generate the first aggregate index; Aggregate the quantified scores corresponding to severity-type index items of the same type in the same target inspection data to generate a second aggregated index; By integrating the text information indexes corresponding to the same target inspection data, a third aggregated index is generated; A target inspection data index is generated based on the first, second, and third aggregated indexes.

[0070] In summary, the image data retrieval system proposed in this application first constructs a medical image database containing multiple target examination data, each including at least one image sequence. Second, it extracts textual description information and image content information for each image sequence, and establishes an image sequence index based on these information. Further, based on the image sequence index corresponding to at least one image sequence included in the same target examination data, it aggregates information according to preset aggregation rules to generate a target examination data index. Finally, in response to a received retrieval request, it performs a retrieval in the medical image database based on the image sequence index and / or the target examination data index, and outputs retrieval results that satisfy the retrieval request. This application divides image content information from multiple dimensions and conducts retrieval based on the textual information and specific image content information of the medical image data. This retrieval method not only eliminates the influence of missing or incorrect image data field information but also supports joint retrieval of image and text information, and provides multi-type feature indexes and retrieval functions related to image content. Based on this, the system can accurately retrieve patient examination data, image sequences, and their text reports, improving the fineness of the retrieval granularity, thereby achieving accurate retrieval and collection of medical imaging data, and more efficiently completing the retrieval of target image data and the construction of datasets.

[0071] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0072] The apparatus of the above embodiments is used to implement the corresponding method in any of the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0073] Corresponding to the above embodiments, this application also proposes an electronic device. (See reference...) Figure 3The diagram below is a block diagram of an electronic device according to some embodiments of this application. It also illustrates a more specific hardware structure of an electronic device provided by an embodiment of this application. The device may include: a processor 310, a memory 320, an input / output interface 330, a communication interface 340, and a bus 350. The processor 310, memory 320, input / output interface 330, and communication interface 340 are interconnected internally via the bus 350.

[0074] The processor 310 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0075] The memory 320 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 320 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 320 and is called and executed by the processor 310.

[0076] Input / output interface 330 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0077] The communication interface 340 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0078] Bus 350 includes a pathway for transmitting information between various components of the device, such as processor 310, memory 320, input / output interface 330, and communication interface 340.

[0079] It should be noted that although the above-described device only shows the processor 310, memory 320, input / output interface 330, communication interface 340, and bus 350, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0080] The electronic devices described in the above embodiments are used to implement the corresponding image data retrieval methods in any of the foregoing embodiments, and have corresponding beneficial effects, which will not be elaborated further here.

[0081] Based on the same concept, corresponding to the image data retrieval method provided in any of the above embodiments, this application also provides a computer-readable storage medium storing a program or instructions, which, when executed by a processor, implements the image data retrieval method as described above.

[0082] The aforementioned computer-readable storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0083] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the corresponding image data retrieval method in any of the foregoing embodiments, and have corresponding beneficial effects, which will not be described in detail here.

[0084] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0085] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.) and includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0086] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. A method for retrieving image data, characterized in that, include: A medical image database is constructed, which contains multiple target examination data, each of which includes at least one image sequence; Extract the text description information and image content information of each image sequence, and build an image sequence index based on the text description information and the image content information; Based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, information is aggregated according to preset aggregation rules to generate a target inspection data index. In response to receiving a search request, the system performs a search in the medical imaging database based on the image sequence index and / or the target examination data index, and outputs search results that satisfy the search request.

2. The image data retrieval method according to claim 1, characterized in that, The step of extracting text description information and image content information for each image sequence, and establishing an image sequence index based on the text description information and the image content information, includes: Obtain an image report associated with the image sequence; A text information index is constructed based on the metadata field information of the image report and / or the image sequence; An image content index is constructed based on the image content information contained in the image sequence; The image sequence index is generated based on the text information index and the image content index.

3. The image data retrieval method according to claim 2, characterized in that, The construction of a text information index based on the metadata field information of the image report and / or the image sequence includes: Extract the first text information from the metadata field information of the image sequence; Extract second text information from the image report associated with the image sequence; The text information index is constructed based on the first text information and / or the second text information.

4. The image data retrieval method according to claim 2, characterized in that, The construction of an image content index based on the image content information contained in the image sequence includes: The image content information is obtained by recognizing the image sequence using an image content recognition model. Generate corresponding index items based on the type of the image content information; The index entries are combined to form the image content index of the image sequence.

5. The image data retrieval method according to claim 4, characterized in that, The image content information includes classification information and severity information; The generation of corresponding index entries based on the type of the image content information includes: Based on the classification information, a classification index item is generated, wherein the classification information is used to characterize whether there is image content of a preset category in the image sequence; Severity index entries are generated based on the severity information, which is used to characterize the degree of presence of specific image content in the image sequence.

6. The image data retrieval method according to claim 5, characterized in that, The aggregation rule includes at least one of the following: For the aforementioned classification index items, classification index items of the same type in the same target inspection data are merged; For the severity index item, aggregate calculation is performed based on the quantified score corresponding to the severity index item; The text information index is then integrated.

7. The image data retrieval method according to claim 6, characterized in that, The process of generating a target inspection data index by aggregating information based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data according to a preset aggregation rule includes: Merge the judgment classification index items of the same type in the same target inspection data to generate the first aggregate index; Aggregate the quantified scores corresponding to the severity index items of the same type in the same target inspection data to generate a second aggregated index; By integrating the text information indexes corresponding to the same target inspection data, a third aggregated index is generated; The target inspection data index is generated based on the first aggregated index, the second aggregated index, and the third aggregated index.

8. An image data retrieval system, characterized in that, include: A database construction module is configured to construct a medical image database containing multiple target examination data, each of which includes at least one image sequence. The sequence index building module is configured to extract text description information and image content information for each image sequence, and to build an image sequence index based on the text description information and the image content information; The data index building module is configured to aggregate information according to preset aggregation rules based on the image sequence index corresponding to at least one image sequence contained in the same target inspection data, and generate the target inspection data index. The data retrieval module is configured to, in response to receiving a retrieval request, perform a retrieval in the medical imaging database based on the image sequence index and / or the target examination data index, and output retrieval results that satisfy the retrieval request.

9. An electronic device, characterized in that, include: A processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the image data retrieval method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the image data retrieval method as described in any one of claims 1 to 7.

Images